Department of Histopathology, GKT Medical School, St. Thomas’ Campus, London, United Kingdom
Correspondence: Jo Spencer, Department of Histopathology, GKT Medical School, St. Thomas’ Campus, Lambeth Palace Rd., London SE1 7EH, UK E-mail: jo.spencer{at}kcl.ac.uk
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
genes in postnatal gut were also studied in the fetus. According to these parameters, fetal and postnatal B cells are unrelated.
Key Words: mucosa • pre-B cells • plasma cells • thymic B cells
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In this study, we have investigated the morphology and phenotype of fetal intestinal B cells and identified two distinct populations. We observed a population of dividing mature B cells with extensive cytoplasmic processes that make T cell contact and bear a striking resemblance to mature thymic B cells. The second, more serosally situated population was identified as pre-B cells, indicating that the fetal intestine is a novel site of B cell development.
Rearrangement of immunoglobulin (Ig) genes during B cell development, activity of endo- and exonucleases, and the random insertion of nucleotides by terminal deoxynucleotidyl transferase result in the generation of VHDJH junctional sequences, which can be considered clonotypic markers for individual B cells. Therefore, we used polymerase chain reaction (PCR), cloning, and sequencing of IgVH genes to confirm the presence of expanding populations of related cells in the fetal intestinal microenvironment, suggested by our immunohistochemical studies.
The question of whether fetal intestinal B cells, the majority of which are CD5+ [4
], and postnatal B cells are related is of interest, as studies in mice have shown that B1 and B2 cells contribute to the intestinal plasma cell population. Approximately 50% of the intestinal plasma cells originate from CD5- B2 cells in the Peyer’s patches. The remaining 50% are derived from a separate lineage of B1 cells, a subset of which is CD5+, which are located in the peritoneal and pleural cavities and from where they seed the gut with antibody-producing cells while maintaining their own population by self renewal [5
, 6
]. The origin of postnatal intestinal plasma cells in man is unclear, and it is not known whether they have a dual origin. The B cell population in the human peritoneum is small, but it does include a population of cells expressing the intestine-specific adhesion molecules 
4ß47 and Eß7 [7
], which therefore have the potential to home to the intestinal mucosa. Although some B cells in human postnatal gut express CD5 when analyzed by flow cytometry [8
], this may be a result of their activation status, as CD5 is an activation antigen in man [9
] and therefore does not necessarily reflect a dual origin.
From previous studies, we have accumulated a unique database of IgVH4–34 genes derived from postnatal intestine that are out-of-frame between VH and JH. As these are unused alleles, the use of D and J by these genes is likely to reflect the factors that bias their use, which operated as the B cells developed. D and J use in VH4–34 genes was therefore compared in fetal intestinal B cells and unselected adult B cells to determine whether fetal and postnatal B cell populations are likely to be related. To address the same question, we investigated light-chain genes in fetal intestine, as studies of light-chain rearrangements in adult intestinal B cells showed that this population had a higher ratio of in-frame:out-of-frame rearrangements than the peripheral B cell population. This suggests that the rearrangement of light chain genes in the precursors of intestinal B cells may be a unique marker for their lineage, which may be common to fetal and adult populations.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Lymphocytes were isolated from the peripheral blood of two healthy adult volunteers, using Ficoll Hypaque (Amersham Pharmacia, Little Chalfont, UK) as controls. In addition, five specimens of normal intestine from patients aged 22.7, 36.7, 66, 77.5, and 80.7 years were also used as controls for analysis of DH and JH bias and light chain conformation.
Immunohistochemistry
Frozen sections (8 µm) were fixed in fresh acetone for 30 min and were then incubated in murine anti-human monoclonal CD20 or IgD (Dako Ltd., Ely, Cambridgeshire, UK) or rabbit anti-human polyclonal IgA or IgG antibodies (Dako Ltd.) for 1 h. Binding was detected using horseradish peroxidase-conjugated rabbit anti-mouse or swine anti-rabbit Ig (Dako Ltd.) as appropriate, followed by diaminobenzidine substrate (Sigma-Aldrich, Poole, Dorset, UK). Sections were counterstained with Mayer’s haematoxylin, dehydrated, and mounted.
Sequential staining was used to analyze cellular proliferation and the relative distribution of B and T cells. The first antibody (to Ki67, IgM, CD20, or CD3, all from Dako Ltd.) was used to stain the sections as described above, but with no counterstain. The second antibody (to IgM, IgD, or CD20) was then applied to the sections, and binding was detected using biotinylated rabbit anti-mouse Ig followed by avidin-alkaline phosphatase conjugate (Dako Ltd.) and fast blue substrate (Sigma-Aldrich). Sections were then washed and mounted in aqueous mountant with no counterstain.
Immunofluorescence was used to study dual antigen expression. Sections were incubated with murine antibodies to IgM (Dako Ltd.) or CD68 (Dako Ltd.) and with a cocktail of rabbit antibodies to 
and light chains (Dako Ltd.). Binding was visualized using fluoresceinated goat anti-mouse Ig (Dako Ltd.) and biotinylated swine anti-rabbit Ig (Dako Ltd.) followed by streptavidin-conjugated rhodamine (Sigma-Aldrich). To detect dual expression of CD20 and IgM, sections were incubated in monoclonal CD20 (Dako Ltd.) and polyclonal rabbit anti-IgM. Binding of CD20 was detected using goat anti-mouse fluoresein isothiocyanate conjugate and anti-IgM using biotinylated swine anti-rabbit Ig, followed by rhodamine conjugated to extravidin.
Reverse transcription and PCR
DNA was isolated from sections of fetal tissue. For two samples, the sections were pooled, and for the remaining two samples, individual sections were processed for use in separate PCR reactions. Sections were incubated directly in proteinase K (Promega, Southampton, UK) to release DNA as previously described [10
]. RNA was prepared from fragments of tissue approximately 2 mm3 using Trizol (Gibco-BRL, Paisley, UK), according to the manufacturer’s instructions. Reverse transcription was carried out using Moloney murine leukemia virus reverse transcriptase (Promega) and oligo(dT) primers (Promega).
Methods for the amplification of IgVH4–34 to JH from DNA samples and IgVH4–34 to C from cDNA samples, using fully nested [10
] and semi-nested methods [11
], respectively, and of V1 and V2 families [12
] have been published previously. PCR products were cloned using the pGEM-T vector kit (Promega) and sequenced on an ABI 377 automatic sequencer using the ABI Prism dye terminator kit (PE Applied Biosystems, Warrington, UK), according to the manufacturer’s protocols. Sequences were analyzed using Genejockey II software and the V base sequencing directory (MRC Centre for Protein Engineering, Cambridge, UK).
Primers for the amplification of V pre-B were designed to span an intron, allowing discrimination between DNA and processed RNA sequences. The primers used were 5'-TGCAGTGGGTTCCATTTCTTCC-3' and 5'-CATGCTGTTTGTCTACTGCACAG-3'. Each primer (100 ng), 200 µM dNTPs (Promega), 1.5 mM MgCl2, and 0.5 U Taq DNA polymerase (Promega) were used in a 50 µl reaction volume. After a hot start of 2 min at 94°C, 30 cycles of 94°C for 30 s, 56°C for 30 s, and 72°C for 1 min were performed. Amplification of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was performed as a control, using the same PCR conditions. The primers used were 5'-GCCTCCTGCACCACCAACTG-3' and 5'-CGACGCCTGCTCACCACCTTC-3'. PCR products were visualized on a 3.5% agarose gel stained with ethidium bromide, and the relative intensity of the bands representing V pre-B DNA and RNA was assessed.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
 View larger version (157K): [in a new window] |
Figure 1. Immunohistochemical visualization of B cell morphology and proliferation and of B cell/T cell contact in human fetal intestine at 20 weeks. (a) CD20+ cells in a primitive follicle. (b) Illustration of morphology of CD20+ cells, which have cytoplasmic-processes, which are often extensive. Frequent contact is made with adjacent cells, and condensation of staining is apparent at points of contact (arrows). (c) Proliferating B cell stained with Ki67 proliferation antigen (brown) and CD20 (blue). (d) Primitive follicle stained with IgD. Note the presence of IgD+ cells with cytoplasmic processes (arrows). (e) IgD-expressing population (blue) includes Ki67+ proliferating cells (arrows). (f) Contact between T cells (brown) and CD20+ cells (blue) can be identified (arrows).
|
 View larger version (32K): [in a new window] |
Figure 2. Identification of pre-B cells in fetal intestine. (a) Expression of IgM (green) and light chain (red) in fetal intestine at 20 weeks. Cells expressing IgM but not light chain are indicated with short arrows. An example of a mature B cell with cytoplasmic processes that coexpresses light chain and IgM is indicated with a long arrow. Inset shows a cell double-stained for light chain and CD68. Cytophilic binding of light chain by CD68+ cells is the likely explanation for the red cells in this figure. (b) Expression of IgM (red) and CD20 (green). Red-only IgM+, CD20- cells are indicated with short arrows. A cell in the same field coexpressing CD20 and IgM is indicated with a long arrow. (c) Identification of V pre-B by PCR. The two arrows indicate the larger product from amplification of DNA and the shorter product representing the expressed gene. Lanes are 1, blood; 2 and 3, fetal intestine at 16 weeks; 4–7, postnatal intestine. V pre-B message is readily identified in fetal intestine at 20 weeks. The weak band identified in blood is consistent with previous observations [13
]. MW, Molecular weight.
|
Human fetal intestine contains pre-B cells
In addition to the large B cells situated beneath the epithelium, a population of small, round IgM+light chain-CD20- cells was present beneath the muscularis (Fig. 2)
. This is the phenotype of pre-B cells, suggesting that local B cell development is taking place. A population of light chain, positive cells that did not express IgM were probably macrophages, which had bound maternally derived IgG, as cells coexpressing CD68 and light chain could be identified (Fig. 2)
. The presence of pre-B cells was confirmed by PCR for V pre-B expression, which clearly demonstrated that there is local surrogate light chain expression in fetal but not postnatal intestine (Fig. 2)
. The low level of V pre-B seen in the blood is consistent with recently published data [13
].
IgVH gene analysis
A total of 91 VH4–34 sequences derived from amplification of DNA from four fetal intestines were analyzed. For two of the intestines, sections were pooled for DNA extraction, and for the remaining two sections were processed individually for separate PCR reactions. Study of sections independently yielded 40 sequences. Of these, 13 were unique rearrangements. The remainder were groups with between two and four members, which shared the same CDR3 sequence but differed slightly in the sequence of the V region. Within three groups of sequences with two, three, and four members, the individual members were derived from different PCR reactions from different tissue sections, consistent with B cell proliferation. The sequences from the group with four members are shown in Figure 3
. The remaining 51 sequences were derived from pooled sections. Of these, three were unique sequences. The remainder could be grouped according to CDR3 sequence, and groups contained between 2 and 15 members. For analysis of D and J use, each group was considered to be a single sequence to avoid distortion of the data.
 View larger version (44K): [in a new window] |
Figure 3. Sequence of IgVH4–34 gene derived from gene-specific PCR of DNA from sections of fetal intestine at 20 weeks. Dashes indicate identity with the germline IgVH4–34 sequence, and base changes are indicated. Section 2 was divided in two halves (a and b), and Sections 3 and 5 refer to subsequent numbered serial sections. Identification of the same CDR3 sequence in IgVH4–34 genes from serial sections confirms that proliferation of cells with rearranged heavy chain is occurring. Base changes from germline could be a result of PCR polymerase error, although the presence of seven base changes in the sequence from Section 3 suggests that somatic hypermutation may occur.
|
In previous studies of D and J segment use by VH4–34 genes in postnatal intestine, we observed 18 IgVH4–34 genes that were out-of-frame between VH and JH. These are considered to be unused alleles that have no selection-introduced biases in D and J use. Therefore, these were used for comparison with the sequences from fetal intestine. Figure 4 illustrates the notable differences between the D and JH repertoire in the fetal intestine and in the unselected repertoire of cells that give rise to plasma cells postnatally. The fetal repertoire is significantly biased toward rearrangement of DHQ52 (P=0.037) and JH4 (P=0.009) when compared with the unselected adult repertoire.
 View larger version (10K): [in a new window] |
Figure 4. Comparison of biases in DH and JH use in IgVH4–34 genes in fetal intestine and out-of-frame alleles from postnatal intestine. Different biases in (a) JH and (b) DH are apparent in fetal B cells (solid bars) and the unselected postnatal repertoire (open bars). These populations are therefore not directly related.
|
1 rearrangement was obtained from a specimen at 16 weeks of gestation (Fig. 5
). It was in-frame, but had two base changes compared with the germline VH4–34 segment, one of which generated a stop codon.
 View larger version (38K): [in a new window] |
Figure 5. A class-switched IgVH4–34DJH6C1 rearrangement isolated from human fetal gut. The top line represents the germline nucleotide sequence of IgVH4–34 gene segment. The second line represents the sequence of an IgVH4–34DJH6C1 rearrangement obtained from a 16-week-old human fetal gut. The UGA translational stop codon located in the FR2 domain is boxed. Dashes indicate identity with the germline VH4–34 gene segment sequence. Nucleotides that are different from the germline IgVH4–34 gene segment are shown. Sequences corresponding to CDR regions are underlined. The JH6 gene segment is shown in italics. *RNA splice site (5' end of C1 gene segment). The 5' IgVH4–34 FR1 and 3' C PCR primer sequences are shown in bold italics.
|
gene analysis light chain genes was compared in fetal intestinal B cells and in CD20- B cells from postnatal intestine. Unlike postnatal intestine, out-of-frame genes of both families were observed in human fetus. In the fetal gut, the ratios were 14:4 in the IgV1J rearrangements and 7:3 in the IgV2J rearrangements. This is similar to postnatal peripheral lymphoid tissue [14
], but different from postnatal, intestinal CD20- B cells (ref. [12
], and L.B. and J.S., unpublished results; Table 1
). |
View this table: [in a new window] |
Table 1. In-Frame:Out-of-Frame Ratios Observed in IgV1J and IgV2J Rearrangements Amplified from Human Fetal Gut B Cells, Postnatal CD20- Areas of Large Bowel Lamina Propria, and Human CD19+ IgM+ Peripheral Blood Lymphocytes (PBLs)
|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In several mouse models, thymic B cells have been shown to play an essential role in negative selection, inducing the elimination of self-reactive CD4+ thymocytes [18 19 20 ]. As extrathymic T cell development occurs in human fetal intestine from 12 weeks of gestation [15 , 21 ], it is possible that as in the thymus, the B cells may play a role in the development and selection of the T cells. However, the fetal intestinal B cell D and J repertoire observed differs from that of postnatal thymic B cells [16 , 17 , 22 ], indicating that they do not share a common origin.
A second population of B cells, which differed from the large B cells in their phenotype, morphology, and location, was observed in fetal intestine and was found to be pre-B cells. This is the first time that the intestine has been identified as a site of B cell development, and the relationship between these pre-B cells and the large, mature cells is unclear. However, as we have shown that the human fetal intestinal B cells are unlikely to be precursors of the adult intestinal plasma cells, intestinal B cell maturation may be a feature of fetal development only. It is interesting that in mice that completely lack Notch-1, a receptor involved in T cell fate specification of T/B precursors [23 ], the precursor cells develop into B cells in the thymus [24 ] but become developmentally arrested in the gut [25 ]. This indicates that the adult murine intestinal microenvironment is not permissive for B cell development.
As the large, mature B cells are considerably more abundant than the small pre-B cells, our analysis of the Ig gene repertoire of fetal intestinal B cells relates to that population. We have shown that their JH and DH repertoire resembles that previously seen in other B cell subsets in the fetus, including the developing B cells in fetal spleen and liver [26
] and the "mature" CD5+ B cells in cord blood [27
]. It is different from the unselected repertoire of plasma cells in postnatal intestine. In addition, a greater proportion of light chain gene rearrangements were found to be out-of-frame in the fetal intestinal B cells rather than in the postnatal, intestinal CD20- plasma cells. Postnatal and fetal B cells are therefore different populations, and there is no evidence for any relationship between them.
As B cell proliferation was seen in fetal intestine, we investigated the possibility that these cells may undergo class-switch recombination. We obtained a class-switched IgVH4–34C1 rearrangement from a 16-week-old fetal intestine. It contained two base changes from the germline in the VH4–34 segment, one of which generated a stop codon that would render the rearrangement unproductive, although the possibility that the base substitution results from a PCR error cannot be excluded. Although class-switch events are rare in the fetus, and there is no evidence for production of IgA, this observation demonstrates that albeit rarely, class-switch recombination can occur in this dividing B cell population in the fetus in the absence of germinal centers or exogenously derived antigen.
Received December 5, 2001; revised March 6, 2002; accepted March 26, 2002.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
J repertoire J. Immunol. 162,2137-2145 and C
repertoires J. Immunol. 150,1348-1357[Abstract]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
